MongoDB 2.0
Good news, everyone. Do you like SQL, or do not like SQL (do you like NoSQL?), But today (oh, already yesterday) came out as PostgreSQL 9.1 (about which, while I am writing this post, someone will probably write too) and MongoDB 2.0 !
In 1.4, two-dimensional geo-indices appeared, in 1.6 — sharding, in 1.8, a bit late — journaling and partial indices ... And what of the jittery appeared in 2.0? The compact command, which can compress only one collection (and not as before - to do repair for the entire database) is not stunning, all sorts of improvements in terms of parallelization and indexes (it is stated that they will now be 25% smaller and 25% faster ) - also boring ...
Perhaps the most interesting thing is that replica set-it became possible to set priorities and tags of their location - well, like, "in which country / in which data center / in which rack" - and create complex rules for these tags, how to save data ( well, up to "each copy of the data should have at least three copies on at least two continents"); This is called the beautiful term Replica Set Data Center Awareness.
')
Also, map / reduce learned how to output data to the sharded collection (and was also optimized and works faster); The $ and operator appeared in the queries; regexps have learned, if the programmer wishes, to match the line breaks with the symbol; The geoindex can be used if several locations are specified for one document at once, as well as for searching inside polygons ...
Something will be in 2.2? .. If we consider that the well-known Sequoia Capital has just invested 20 million dollars in 10gen (and they have already had 10 million more), it looks like it won't be boring.
In 1.4, two-dimensional geo-indices appeared, in 1.6 — sharding, in 1.8, a bit late — journaling and partial indices ... And what of the jittery appeared in 2.0? The compact command, which can compress only one collection (and not as before - to do repair for the entire database) is not stunning, all sorts of improvements in terms of parallelization and indexes (it is stated that they will now be 25% smaller and 25% faster ) - also boring ...
Perhaps the most interesting thing is that replica set-it became possible to set priorities and tags of their location - well, like, "in which country / in which data center / in which rack" - and create complex rules for these tags, how to save data ( well, up to "each copy of the data should have at least three copies on at least two continents"); This is called the beautiful term Replica Set Data Center Awareness.
')
Also, map / reduce learned how to output data to the sharded collection (and was also optimized and works faster); The $ and operator appeared in the queries; regexps have learned, if the programmer wishes, to match the line breaks with the symbol; The geoindex can be used if several locations are specified for one document at once, as well as for searching inside polygons ...
Something will be in 2.2? .. If we consider that the well-known Sequoia Capital has just invested 20 million dollars in 10gen (and they have already had 10 million more), it looks like it won't be boring.
Source: https://habr.com/ru/post/128284/
All Articles